Research on N-gram-based malicious code feature extraction algorithm

The amount of computer virus is on the increase since its first appearance and has posed serious security threats to the computer systems. Most of the current anti-virus systems attempt to detect these new malicious programs through heuristics scheme, but this costs a lot and is often ineffective. In this paper, an N-gram-based malicious code feature extraction algorithm, based on statistical language model, is presented. Through this algorithm, the N-gram features of the sample set can be extracted and the features of the malicious code can be obtained exactly. Compared with the traditional feature code-based approaches, our approach has higher detection rates for new malicious codes.